Abstract:
Systems and methods for simultaneously capturing live audio and video feeds and reducing motor/rotor noise associated with the video feed for a multirotor unmanned aerial vehicle (UAV). Noise reduction occurs in real time or near real time as one or more frequencies associated with a motor, rotor, or attitude of the UAV are recognized, subtracted and/or dithered from the audio feed, resulting in one or more correction signals. The one or more correction signals may be dynamically summed in order to generate a corrected audio feed for transmission.
Abstract:
A system and related method for tracking a moving subject via a ground-based or airborne peripheral device includes plugging an intelligent camera device to the unmanned vehicle, establishing a peripheral link by which the intelligent camera device can assume control of the unmanned vehicle's control systems. Based on analysis of images captured by the intelligent camera device, in addition to position data associated with the unmanned vehicle or with the subject, the intelligent camera device may autonomously maneuver the unmanned vehicle to track or follow the subject while maintaining the moving subject in a consistent framing orientation.
Abstract:
Systems and methods for tracking a subject using an unmanned aerial vehicle (UAV) are disclosed. An unmanned aerial vehicle (UAV) includes an onboard camera to capture/stream multiple images. An onboard visual recognition module isolates image elements by analyzing the captured images, and determines whether the image elements correspond or do not correspond to the subject. Current image elements corresponding to a subject are stored in a positive database, and non-corresponding current image elements are stored in a negative database. An onboard subject tracking module defines the chosen subject, determines attributes of the subject based on comparisons of image elements, and follows the subject by directing the attitude control system to adjust the velocity and orientation of the UAV based on the determined attributes.
Abstract:
An action camera system for an unmanned aerial vehicle (UAV) selects a target based on a reference image captured by an onboard camera. Image processing determines a desired orientation of the target to the UAV, by which the UAV can track the target and provide streaming video images from the desired orientation. Image processing establishes a visual lock on the target and controls the UAV to maintain the desired orientation while capturing streaming images. Additional position data provided by sensors aboard a smartphone carried by the target enhances the tracking ability of the action camera system and enables predictive analysis of the target's future position. The action camera system may additionally provide preselected modes of operation that control UAV movement and image capture depending on the user's desired objectives.
Abstract:
Systems and methods for tracking a subject using an unmanned aerial vehicle (UAV) are disclosed. An unmanned aerial vehicle (UAV) includes an onboard camera to capture/stream multiple images. An onboard visual recognition module isolates image elements by analyzing the captured images, and determines whether the image elements correspond or do not correspond to the subject. Current image elements corresponding to a subject are stored in a positive database, and non-corresponding current image elements are stored in a negative database. An onboard subject tracking module defines the chosen subject, determines attributes of the subject based on comparisons of image elements, and follows the subject by directing the attitude control system to adjust the velocity and orientation of the UAV based on the determined attributes.